Description of replication files for "Testing Rank Similarity," by Brigham Frandsen and Lars Lefgren

star_example_final.do: Stata do-file which produces the empirical results in the paper

star_example_final.log: log file produced by the above .do-file

STAR_Students.dta: Stata dataset containing treatment assignment, actual enrollments, outcomes, and covariates on students enrolled in the the Tennessee STAR experiment. This dataset is a simple conversion to Stata format of the publicly available dataset STAR_Students.tab located at https://dataverse.harvard.edu/dataset.xhtml?persistentId=hdl:1902.1/10766. A detailed description of the data can be found at the dataverse URL in the previous sentence, but the following are the variables used in the analysis for this paper:
gkfreelunch: indicator for the student being eligible for free or reduced-priced lunch (an indicator of low parental income)
gkclasstype: indicator for treatment assignment. This variable equals 1 when the student was assigned to a small class in kindergarten
g1tmathss: score on 1st grade math assessment; our main outcome variable


ranktest.ado: Stata .ado-file which defines the Stata command "ranktest" which implements the testing procedure

abadiels.ado: Stata .ado-file which defines a subroutine called by "ranktest"

control_high_low_comp_STAR.pdf(gph): figure produced by star_example_final.do, which corresponds to Figure 2 in the paper 

treat_control_low_comp_STAR.pdf(gph): figure produced by star_example_final.do, which corresponds to Figure 3 in the paper 